Metagenomics and biogeochemistry

1
Metagenomics and biogeochemistry How do
microorganism-driven geochemical cycles affect
structure and function of ecosystems? How do we
assess structure and function of ecosystems? How
about starting by relating microbial assemblage
composition to biogeochemical parameters and
functions? Can we find predictable
relationships? Patterns and scales of
variability? Is metagenomics (e.g. shotgun or
large-insert libraries) the best way to assess
microbial assemblage composition for such
studies? Are there faster and cheaper ways that
permit analysis of many samples?
2
Amplified Ribosomal Intergenic Spacer Analysis
(ARISA)
For microbial community fingerprints with high
phylogenetic resolution
Start with DNA extracted from a mixed
community. PCR spans rRNA operon, 16S to 23S
genes. One tagged primer.
PCR
Fragment Size
Fluorescence
Fragment analysis. Smallest detectable peak 0.1
of total
Shows exact sizes. Each peak Operational
Taxonomic Unit. Data based, not gel-based.
Ref Fisher and Triplett 1999
16S-23S clone libraries to identify most peaks
Brown, Hewson, Schwalbach Fuhrman, Envir.
Microbiol 2005
3
16S-ITS Clone Library permits ID from ARISA.
Example USC Microbial Observatory 512 clones
cover 94 of ARISA peaks
Brown et al. 2005, Envir Microbiol.
4
Quantitation from PCR-based Fingerprinting?
Real comparison Prochlorococcus, ARISA vs flow
cytometry counts
San Pedro Ocean Time Series 4 year dataset
Note we use a highly standardized assay, with
eukaryotes removed, and measured amounts of DNA
R20.86
Flow cytometric counts
area from ARISA
Fingerprint area is remarkably proportional to
counts. Also, SAR11 clones are close to
cells.
Brown, Hewson, Schwalbach Fuhrman, Envir.
Microbiol 2005
5
Replicate 20L samples have very similar ARISA
fingerprints
7 samples from each of 2 North Pacific Gyre
Stations
Compares OTU proportions OTU
Presence/absence only
Hewson et al. Aquat Microb Ecol 2006
6
What is an ARISA OTU? Phylogenetic resolution is
about 98 16S rRNA similarity - comparable to
species level
Easily determined difference
Brown et al, Env Microbiol 2005
7
Near-surface SAR11 subclades as determined by ITS
sequences and lengths
8
Temporal Variability in Bacterioplankton
Communities How fast do communities change?
San Pedro Ocean Time Series
USC Microbial Observatory Measured Microbial and
Oceanographic properties monthly since 2000, at
depths to 880 m Also, daily measurements near
USC Wrigley Marine Science Center on Catalina -
open water accessible daily by small boat Follow
taxa by ARISA to look for temporal diversity
patterns
9
Relative stability over days at one location
(open water, Catalina)
Abundant taxa vary little
Graphs all OTU over 6 days
g
SAR 11
Actinobact
date
Rarer taxa can vary more
Not just noise in measurement
a
g
Prochlorococcus
.
Rarest detectable taxa
CFB
g
SAR 11
10
Monthly observations at SPOTS over 4 years showed
some taxa clearly had repeatable seasonal
patterns. How about the bacterial community in
general?
Brown et al. 2005
11
Predictable Annual Bacterial Community Reassembly
Fuhrman et al., PNAS 2006
with Shahid Naeem
171 taxa followed by ARISA over 4.5 years DFA
scores reflect quantitative distribution of taxa
via ARISA
12
DFA showed some subsets of bacterial taxa could
predict the month of sampling with 100 accuracy.
Multiple Regression with environmental
parameters was highly significant (r2 0.7)
implies predictability of bacterial communities
even in an open marine system. Different subsets
of taxa were predictable from different
parameters implies niches. Highly repeatable
and predictable patterns imply little functional
redundancy, contrary to common expectation for
bacteria. This refers to combinations of
functions in a particular taxon. Note- Not all
taxa were included in the predictable subsets,
but most were.
Significant Parameters in MRA temperature,
salinity, nitrite, nitrate,
silicate, oxygen, bacterial and viral
abundances, bacterial production via leucine and
thymidine incorporation, chlorophyll,
phaeopigments ARISA richness
13
The taxa that had significant multiple regression
coefficients were affected by different
parameters many controlling factors, and
different taxa controlled differently (niches).
14
Biogeography on a Global Scale

• Global survey of bacterioplankton at numerous
  sites in 3 ocean basins, under Arctic ice cap,
  and near Antarctica

15
Global Diversity Measurements via
ARISA Assemblages clearly vary
Things change
16
Bacterioplankton Biogeography

• LATITUDINAL GRADIENT OF RICHNESS
• ARISA measured the same way from 78 samples
  collected in all seasons and both hemispheres
  over 10 years (opportunistic sampling)
• Diversity generally highest at low latitudes,
  lowest in polar environments like animals and
  plants (in every general biology textbook)
• Contrasts sharply with results reported for
  protists

plt0.005
Highly significant (plt0.005) as linear
regression, rank correlation, or with potential
outliers removed
17
Regional Diversity Patterns Bacterial Community
Similarity (via ARISA) vs Distance NEAR-SURFACE
samples
Mixing curve between Pacific and Indian Basins?
Hewson et al 2006 Mar. Ecol. Prog. Ser.
18
Deep-Sea (500-3000 m depth) patterns differ with
locations and depth. Cause(s) unknown North
Atlantic 1000m depth samples were in vicinity of
Amazon Plume
Pacific

Pacific

Hewson et al. 2006 Limnol. Oceanogr.
19
Go beyond just observing nature - EXPERIMENTATION
Example - What does proteorhodopsin do? Does it
provide much energy, and help microbial growth,
as many assume? Genomics alone cant
answer. Schwalbach et al. (2005 Aquat. Microb.
Ecol. 39 235 ) did light/dark experiments with
oceanic plankton. Water collected from
oligotrophic and mesotrophic Pacific Ocean
locations, collected and stored in natural light
or total darkness for 5-10 days. Bacterial
assemblages monitored by the ARISA
whole-community fingerprinting approach
20
EXPERIMENTAL TEST of Significance of
Phototrophy. Light Removal Experiments focus on
Bacterial Groups that are supposed to have
Proteorhodopsin
Incubate bacteria in Light or Dark for 5-10 days
Dark 24hr
DAPI
Cell Abundances Monitored over time
Mesocosms (2x20L)
Collect Cells After 5-10 days
P3
DNA Extraction
ITS Clone Library Construction
Bacterial Community Composition
rDNA
PCR
PCR
Clone Sequence
16s
ITS
23s
DNA
16S-ITS-23S
ARISA Delineate 98 16s rDNA
ABI 377XL
Database of ARISA OTU Identities
Automated Ribosomal Intergenic Spacer Analysis
21
Schwalbach et al Aquat Microb Ecol 2005
Light Removal Experiments, 5-10 days darkness
Histogram summarizing magnitude of change in
individual taxa, light vs dark treatments Most
taxa were NOT affected by light removal
of OTU
Light preference
Dark preference
Cyanobacteria Phytoplankton exhibited
consistent preference for light treatments Mixed
Responses, mostly dark preference, in ALL OTHER
phototrophic groups (e.g. SAR11, SAR86, CFB,
Roseobacter)
Number of taxa displaying response, ALL
experiments
22
Conclusions of Schwalbach et al (2005) Most
taxa (including presumed PR-containing and
bacteriochlorophyll a containing groups) do not
decline significantly in extended darkness,
unlike cyanobacteria. In fact, most bacterial
groups did no differently or much better in
extended darkness than in normal light.
Suggests no clear direct benefit from light for
most organisms. But some organisms do benefit.
23
Even the one pure culture that contains
proteorhodopsin grows no better in the light than
in the dark Pelagibacter, in SAR11 cluster
The Pelagibacter proteorhodopsin functions as a
light-dependent proton pump. The gene is
expressed by cells grown in either diurnal light
or in darkness, and there is no difference
between the growth rates or cell yields of
cultures grown in light or darkness.
Giovannoni et al. Nature
2005
24
Acknowledgements
NSF, esp. Microbial Observatories Program USC
Wrigley Institute Dave Caron Mark Brown Ian
Hewson Mike Schwalbach Josh Steele Anand
Patel Shahid Naeem Tony Michaels Doug
Capone Ximena Hernandez R/V Kilo Moana R/V
Seawatch Ajit Subramaniam Burt Jones
25
Other Issues Quantitation from Environmental
Genomic Data Accurate prediction of
biogeochemical (or any other) function from
genes. Genome Rot, Multifunctional genes, e.g.
generic reductases. More important with
slow-growing organisms and streamlined
genomes?
26
Quantitation Issues/Problems PCR Clone Libraries
Copy number bias mentioned yesterday. Primer
Choice/Bias, Extension Bias? Yes, but how
bad? Example Marine Archaea compared to
Bacteria. DISTANT Fuhrman et al. (1992) used
universal primers, found 5 of 7 clones from 500 m
were Crenarchaeota. DeLong (1992) used archaeal
primers with surface waters only, and RNA
hybridization to compare to Bacteria. Archaea
lt2. Fuhrman and Davis (1997, univ. primers)
Archaea were 1/3 of clones from 500 m 3000 m,
Atlantic and Pacific FISH results Fuhrman and
Ouverney 1998, Archaea to 40 at 600 m in
Pacific, 60 at 200 m in Mediterranean. Karner et
al. (2001) Archaea 30 below 200m at HOT
over gt 1 year. Note If QPCR shows doubling each
cycle and if not at the saturation point,
anything primed OK should quantify OK
27
Metagenomics BIAS? Missing rRNA genes from
large-insert library
All BLAST hits-
SSUrRNA Genes- Presence/absence
DeLong et al. Science, 2006
SAR11